Cable television networks, including community antenna television (CATV), hybrid fiber-coaxial (HFC), and fiber networks, have been in widespread use for many years and are extensive. The extensive and complex cable networks are often difficult for a cable operator to manage and monitor. A typical cable network generally contains a headend which is usually connected to several nodes which provide bi-directional content to a cable modem termination system (CMTS). In many instances, several nodes may serve a particular area of a town or city. The CMTS contains several receivers, and each receiver connects to several modems of many subscribers. For instance, a single receiver may be connected to hundreds of modems at customer premises. Data may be transmitted downstream to the modems on different frequency bands. The modems communicate to the CMTS via upstream communications on a dedicated frequency band, referred to as a return band.
Cable networks are also increasingly carrying signals, which require a high quality and reliability of service, such as Voice over IP (VoIP) communications. Any disruption of voice or data traffic is a great inconvenience and often unacceptable to a customer. Various factors may affect the quality of service, including the quality of the upstream channels. One factor that affects the quality of upstream communications is the presence of up-stream channel impairments, such as micro-reflections (MRs) of communication signals, group delay variation (GDV), and amplitude distortion (AD).
AD is an undesirable variation in the channel's amplitude response. Common forms of AD include tilt, ripple, and roll-off. A common cause of AD is upper return band-edge carriers, aggravated by long reaches of a cable network plant. The long reaches accumulate diplex filters from devices including amplifiers and in-line equalizers. While individually contributing small attenuation versus frequency, the accumulated diplex filters can create appreciable response variation. In a QAM constellation, the amplitude roll-off causes the symbols to spread in a pattern similar in appearance to Additive White Gaussian Noise (AWGN) and causes received symbols to cross decision boundaries, resulting in errors.
GDV is an undesirable variation in the communication channel's phase response, resulting in distortion of the digital signal phase, or a variation in the propagation of frequency components of the signal across the channel. As is the case for AD, one major cause of GDV in the plant is upper-band-edge operation, combined with long reaches of cable network plant. The reasoning is the same as in the AD case. Note that filtering functions typically induce nonlinear phase responses as the band edges are approached, so the combination of AD and GDV in the same band region is perfectly expected, with the understanding that diplex filtering is the cause. Different filter functions induce different GDV responses, in a similar manner that different filter functions induce different stop-band characteristics. It is typical that the sharper the roll-off, such as would be the case for long cascades, the worse the GDV will be. In a QAM constellation, GDV causes the symbols to spread in a pattern similar to AWGN and AD and causes received symbols to cross decision boundaries, resulting in errors. 16-QAM is less sensitive to GDV than 64-QAM because of reduced decision boundary size of 64-QAM.
As seen by a receiver, a MR is a copy of the transmitted signal, arriving late and with reduced amplitude. The result of the additional copy is the typically seen by end users as image ghosting in analog video reception, whereas for digital communications the result is inter-symbol interference (ISI). MR sources are composed of pairs of hybrid fiber-coaxial (HFC) components separated by a distance of cable. The HFC components, also referred to as cable network components, facilitate the propagation of signal copies in a variety of ways including return loss, isolation, mixing, and combining. For instance, the MR may arise if a length of cable separates two devices with poor return loss, acting as signal reflectors. The reflector return loss and the loss between the reflectors determine the amplitude of the MR. Any HFC component, for instance a cable modem (CM), has the potential to act as a signal reflector. Note that the CM typically has as a design limit of 6 dB return loss, meaning it may reflect up to 25% of its incident power. In the cable network plant, components other then the CM typically reflect a lower percentage of incident power because the design limits are typically significantly better. However, as the cable network plant ages and elements that contribute to good RF matching degrade, for instance connectors, cable, splitters, and interfaces on printed circuit boards (PCBs), the reflected percentage of incident power increases.
These upstream channel impairments are known to be mitigated by the fundamental digital communications receiver function of equalization. During equalization, an equalizer generates coefficient information that is used to create an equalizing filter, with an inverse channel response, canceling distortion in the channel caused by the upstream channel impairments. The equalization coefficients in Data Over Cable Service Interface Specification (DOCSIS) 2.0 and DOCSIS 3.0 are 24 symbol-spaced coefficients (also referred to as taps). Equalization is part of virtually all modern telecommunications platforms, and is instrumental in proper return operation for all DOCSIS systems.
In order to offer higher data rates to subscribers in the competitive world of high-speed data and Internet access, operators must take advantage of the throughput benefits gained from leveraging more complex digital modulation schemes, such as 32-QAM and 64-QAM. Use of 32-QAM allows, for example, a 20 Mbps 16-QAM upstream to become a 25 Mbps upstream. On the other hand, for 64-QAM, it allows a 16-QAM, 20 Mbps upstream channel to become a 30 Mbps channel. This represents a 25-50% throughput improvement. Unfortunately, channels using these digital modulation schemes are also considerably more sensitive to digital communication channel impairments, including the upstream impairments described above, than the 16-QAM channels they are often replacing in the return band.
Given the potential problems that can be caused by the upstream impairments, upstream channels are one of the most challenging digital communication channels to manage and fully exploit. Operators prefer to ensure that capacity associated with the upstream channel, or as much of the capacity as possible, is realized for services and revenue. To do so requires a thorough understanding of a diverse set of HFC and digital communications variables. More importantly, variables that did not matter very much for 16-QAM operation now become not just relevant, but critical to understand for successful deployment of 64-QAM, and to a lesser extent, 32-QAM. Accurately diagnosing upstream issues typically requires technicians or engineers to be at multiple locations within a HFC plant and simultaneously inject test signals at the suspected device locations. This diagnostic process requires extensive manual effort, often requiring rolling trucks to remote locations within a plant or specialized test equipment. The diagnostic process is also time consuming and costly.